International Journal of Applied Science and Technology

Vol. 4 No. 2; March 2014

A Novel Numerical Approach to the Dynamics Analysis of Marine Cables

Iordan Matulea
Drago tefan
Department of Mechanical Engineering
Dunrea de Jos University
800008 Domneasca Street No.47, Galai,
Romania
Bogdan Vladescu
Cristian Barbu
Military Technical Academy
39-49 George Cosbuc Ave.Sector 5, 050141
Bucharest, Romania
A. M. Gonalves-Coelho
UNIDEMI
Faculdade de Cincias e Tecnologia
Universidade Nova de Lisboa
Campus da Caparica
2829-516 Caparica, Portugal

Abstract
Marine cables are exposed to the action of water in relative motion, and hence the need to predict the dynamic
behavior of these complex mechanical systems under the typical working conditions since the outset of the
corresponding design processes. This paper describes the basic principles that can be used in a computer
program meant to support the design of marine cables that should work under a combination of static and
dynamic loadings. A unique mathematical model and a matching numerical method based on finite differences
are used, first for determining the static equilibrium configuration of the riser, and then for findings its dynamic
response around the formerly computed static configuration. A computer program was specially developed for
implementing the proposed model, and a step-by-step method for the design of marine cables similar to the ones
in operation at the Black Sea are herein presented. The conclusion is that this new model provides a coherent and
efficient means to analyze the dynamic behavior of marine cables, as required during the design process.

Key Words: marine cables; dynamic behavior; finite differences; optimization

1. Introduction
The main objective of this study is to deploy a coherent and efficient mathematical model for the cable
mechanics. Beginning from this of this model will be determinate the cable/chains mechanics for different
technical situations. Starting from this model, the cable behavior for different technical situations is analyzed, and
the corresponding original computer programs were used to determine the cables mechanics. The marine
cable/chains utilization presents actually a great diversity including:
1. Moored systems: vessels, offshore exploration units, surface or underwater body, floating dock, other floating
structures; 2. Towed systems: surface vessels, underwater body, fishing nets.
Problem Formulation
A flexible towing/mooring cable presents a difficult problem in dynamic analysis so looking the mathematical
model formulation but especially concerning the numerical computation.
178

 Center for Promoting Ideas, USA

www.ijastnet.com

By making key assumptions is possible to obtain a satisfactory accuracy of mathematical model and of
computation results respectively. Several mathematical models are used to describe the dynamical behavior of
cable. In the present time our experience is confirmed relative to two principal methods, which are applied in the
design of moored/towed systems. The first approach taken for moored/towed systems prediction has been to
represent the continuous cable with a series of rigid segments having the mass-elastic characteristic concentrated
at discrete points along the cable. It is a development of the lumped-mass method [7]. The equations of motion are
obtained by expressing the condition of dynamical equilibrium DAlembert principle- for each cable element or
these equations of motion are derived by Lagranges equations. One can mention the analyses of Walton &
Polachek were based on the use of a lumped characteristics model [12]. In 1980, Nakajima et al. developed a new
method based on a lumped mass model that can be used for the complete analysis both static and dynamic
of cable mechanics [10].
The second approach taken for moored/towed systems prediction has been to represent the marine cable as a
continuous medium having the mass-elastic characteristic distributed along the cable. The cable motion is
described by a system of partial differential equations. Wicker, L.E. obtained a solution for these equations using
the method of characteristic based on the partial differential equations of the cable motion in a bi-dimensional
space [1]. Bernitsas, M. M. developed a study for a three dimensional nonlinear large deflection model for
dynamic behavior of risers, pipelines and cables [4]. In addition, several researchers have developed mathematical
models to take in account the bending stiffness and torsion stiffness [5, 6], as well as more accurate expressions
for the hydrodynamic forces and for the mass distribution [6]. Some other methods for the mooring/towing cables
design were developed [2, 3, 5, 6, 7, 10, 11]. Several studies on cable systems have been made by our R&D
group. The present paper reports the basic approach that has been made on modeling a mooring/towing system, on
deriving the equations for the marine cable mechanics, and on the introducing the necessary engineering
assumptions to achieve practical solutions.
Moored/Towed Systems
The approach taken for moored/towed system has been to represents the continuous cable composed of elastic
segments ds having mass and elastic characteristic distributed along the elements. As such the mechanical
model of cable dynamics approximates the set of exact partial differential equations with another set of non-linear
ordinary differential equations because they are amenable to direct solution by computation. The assumptions can
be summarized as follow:
1. The cable is represented by n rigid segments;
2. The mass-elastic characteristic of cable is distributed along the each segment;
3. The elongation of the cable is considered;
4. Accelerations arising from changes in the towing ships turning rate or speed are assumed to be
negligible;
5. Accelerations due to the products of angular rates-gyroscopic effect- are assumed to be negligible;
6. Dynamic coupling between model segment due to inertial effects can be neglected;
7. The towline length can be constant or variable respectively.
The hydrodynamic forces acting on each element are determinates by using the relative velocity distribution along
each element. These forces are reduced at the discrete points of cable. The hydrodynamic forces are nonlinear
functions of the relative water velocities of each segment.

2. The Dynamics of Marine Cable in a Bi-Dimensional Space

In several cases, both the ends of the cable and the water flow are located in the same vertical plane (Figure 1).
Also the cable excitation is made in the same plane. In those cases, the cable motion is performed in the same
vertical plane. As a general feature of our approach we have developed a mathematical model for dynamic
analysis of cable motion from which, by particularizing, one can deducts the model of marine cable equilibrium
configuration (Figure 1). The marine cable equilibrium configuration is needed for the complete analysis of
marine cable motion, because it determines the initial configuration of the cable, which is considered an
equilibrium configuration. A broad model has been developed to support the prediction of the cable motion in an
offshore operation condition. The model is composed by the differential equations of the cable motion, and the
related methods can be used in several specific situations:
179

International Journal of Applied Science and Technology

Vol. 4 No. 2; March 2014

1. Single or compound moored systems (Composed by cables, chains, underwater or surface bodies, buoys,
weights and docks.);
2. Distinct boundary conditions at both the lower and the upper ends;
3. Different loads, both concentrated and/or distributed;
4. Different continuity condition in the points where bodies are located, or in the points where distinct cables
are coupled;
5. Distinct bi-dimensional and three-dimensional geometric configurations;
6. Diverse water depth;
7. Diverse seabed topography;
8. Single or compound mooring/towing systems.
The solutions can be obtained in two different cases:
1. The boundary conditions are imposed at both ends of the cable. Only a part of the configuration parameters
is known, and the remaining parameters are determined through computation;
2. The boundary conditions are imposed only at one of the cable ends.
All the parameters must be known in the cable end where the computation begins. The conditions that are
required for stopping the computation are imposed in the other cable end.
Kinematics Relationships
Taking into account the fact that the differential equations of cable motion will be solved by integrating step by
step on time, the relationship between speed, velocity and displecemts must be expressed as it follows.
If the index of cable point M is noted by and the index of cable point M 1 is noted by +1 (Figure 2a) then
we can writte.
For the current point of cable, M, we can note:
a. Position vector of point M in the coordinates system xOz:
( r , t ) ( x , t

z , t ) T ; ( r , o ) ( x , o

z ,o ) T

(1)

where (r ,t ) defines the current position of point M at the instant t in the coordinate system xOz, and (r ,o ) defines
the initial position of point M relative to the coordinate system xOz; x , t , z , t are the coordinates of the current
point, M, in the coordinate system xOz at the current time, t; and x ,o , z ,o are the coordinates of the current
point, M, in the coordinate system xOz, at the initial time, t o , as found during the previous static analysis.
b. Velocity vector of point M at time t+t in the coordinate system xOz:
( r , t t ) ( x , t t z , t t ) T

(2)

and velocity vector at time t:

( r , t ) ( x , t z , t ) T

(3)

where t is the time step for the displacement of point M.

The components of the velocity vector at time t+t, can be expressed as:
x , t t x , t

x , t t x , t
t

; z , t t z , t

z , t t z , t
t

(4)
where x , t t and z , t t are the coordinates of current point M, at time t+t.
c. The acceleration vector of point M in the coordinate system xOz, at time t+t , can be expressed as
(r, t t ) (x , t t z , t t ) T

(5a)

The same vector at time t can be expressed as:

(r , t ) (x , t z, t ) T

180

(5b)

 Center for Promoting Ideas, USA

www.ijastnet.com

The components of the vector (5a) can be expressed as:

x , t t x , t

x , t t x , t
t

; z , t t z , t

z , t t z , t

(5c)

The projections of the velocity and of the acceleration of the midle point of the element ds (Figure 2), can be
written as
x 1 x
z
z
; z i 1
2
2
x
x
z
z
v x (z i ) x i 1
; v z (z i ) z i 1
2
2
x 1 x
z1 z
xi
; zi
2
2

(6a)

xi

(a c ) t (x i, t

i=1,...,n;

(6b)
(6c)

zi, t ) T

(6d)

=1,...,n; where n is the number of riser elements.

The velocity of the riser relative to the water in the coordinate system xOz (see Figure 2b) can be written as:
v r (z i ) [ [ v c (z i ) sin ] v x (z i )] i [ [ v c (z i ) cos ] v z (z i )]k
(7a)
where v c (z i ) is the water current velocity in the midle point of the current element ds, (Figure 2), and
v x (z i ), v z (z i ) are the components of the velocity vector (v(z i )) at the midle point of element ds relative to the
inertial coordinates system xOz.
One should notice that:
( v(z i )) ( v x (z i )

v z (z i )) T

; ( v c (z i )) ( v cx (z i ) v cz (z i )) T

(7b)

Because the parameters of the next riser point are unknown during the iterative computation of the equilibrium
riser configuration, the expression of the relative velocity is written as:
v r (z ) [v c (z ) sin ] i [ v c (z ) cos ] k
(8)
The relative velocity unit vector, corresponding to the middle point of element ds can be expressed as:
u (z i )

v r (z i )
u x (z i ) i u z (z i ) k
v r (z i )

(9a)

and the relative velocity unit vector in the current riser point is:
u (z )

v r (z )
v r (z )

u x (z ) i u z (z ) k

(9b)

Where
u x (z i ) [ v c (z i ) sin (z i ) v x (z i )] / v r (z i )

(10a)
u z (z i ) [ v c (z i ) cos (z i ) v z (z i )] / v r (z i )

For determining the equilibrium configuration, we will use:

u x (z ) [ v c (z ) sin (z )] / v r (z )

(10b)
u z (z ) [ v c (z ) cos (z )] / v r (z )

The hydrodynamic forces acting on the riser unit length (Figure 2a) are expressed in the coordinate system of
unit vectors , n, which expressions are:
Tangent unit vector, : xi zk cos i sin k
Normal unit vector, n: n n x i n z k sin i cos k
181

International Journal of Applied Science and Technology

In the system of unit vectors
cable unit length [1]:

Vol. 4 No. 2; March 2014

, n, we have the following expressions for the hydrodynamic forces acting on the

1
F w d e c v 2r [ 0.083 cos ( v r , ) 0.035 cos 2 ( v r , )]
2

(11)
1
Fn w d e c n v 2r sin ( v r , ) sin ( v r , )
2

where cos(vr, ) = u , w is the water density, c

cn are the hydrodynamic coefficients for the bare cable [1],
and d e is the external diameter of the cable hosepipe.
One can neglect the effect of the frictional forces when the water flow velocity is low. According the Podes
testing data [1], usually one can neglect the tangential component of the hydrodynamic force. The hydrodynamic
forces for the cable element can be expressed in the inertial coordinates system xOz using the matrix notation.
Denoting by (F) the vector of the hydrodynamic forces in the system of unit vectors , n, and by (f) the vector of
the same forces in the coordinate system xOz, we can notice that
Fn ) T ; ( f ) ( f x

( F) ( F

f z )T

(12)

And

(f ) R (F)

(13)

where [R] is the transformation matrix between the system of unit vectors

[R ] x
z

n x cos
sin

n z sin cos

, n, and the coordinate system xOz:

(14)

For the tension at both ends of the riser element ds, we have:
(15)

(T) [R](TL )

where:
(T ) ( Tx

Tz ) T

(16)

(TL ) ( T

0) T

(17)

At the lower end of element ds, there is

(T1 ) R dR (T1L )

(18)

where d[R is the total differential of the matrix [R

neglected. In addition:
(T1L ) (T d T

the second and higher-order infinitesimals being

0) T

(19)

To apply the DAlemberts principle we need to know the vector of inertial forces. The mapping of the vectors of
inertial force between the coordinate system xOz as (Fi) and the system of versors , n, i.e. (FiL), is given by:
(Fi ) (Fin ) R (F )
(20)
Where
(Fi ) (Fix Fiz ) T ; ( Fin ) ( Fin , x Fin , z ) T ;

Fin , x c x i ; Fin ,z czi

(21)

with c being the mass per unit length of the cable.

The vector of added mass per unit lenght of the cable is
(F ) (F F n ) T ,

which elements in the the system of unit vectors

F a ; F n n a n

182

(22)
, n are given by
(23)

 Center for Promoting Ideas, USA

where the acceleration projections in the the system of unit vectors
1
n p w c in d e2 ;
2

www.ijastnet.com
, n, are a and an, and

1
w c i d e2 ,
2

(24)

w being the mass density of water, d e the external diameter of the cable, and c in , c i the inertial coefficients that
are determined through experimental tests.
Based on the expression
(a ) t R T (a c ) t

(25)

where the (a)t can be expressed by the relationship:

(a ) t (a

a n ) Tt

(26)

one can write:

a x xi, t zzi, t cos ( t )x i, t sin ( t )zi, t

(27)
a n n x x i, t n zzi, t sin (t )x i, t cos ( t )zi, t

Now, applying the DAlemberts principle, the dynamic equilibrium equation of element ds can be written as:
(T ) (T1 ) ( f )ds ( P)ds (Fi )ds 0
(28)
where ds is the initial length of the cable element. The weight vector, (P), for the unit length of the cable is:
(P) ( 0

q) T

(29)

where q is the weight in water per unit cable length, and (T), (T1) are the cable tension vectors at both ends of the
cable element. Thus, the differential equations of cable motion can be written as:
d a2 e2
dT
( a 1 e 1 )c T ;
(
)c T ;
ds
ds
T
T
dx
dz
dS
T
cT;
c T cos ;
c T sin ; c T 1
ds
ds
ds
EA

(30)

where:
a 1 q sin c (x i cos zi sin ); a 2 q cos c (x i sin zi cos )
e1 F a ; e 2 Fn n a n

After the integration of the Eq. (30) one can integrate the equations
x

dx
dz
dx
dz
; z ; x
; z
dt
dt
dt
dt

(31)

The Static Equilibrium Configuration

In the specific case of the equilibrium configuration, by particularizing Eq. (30)
a 1 q sin ; a 2 q cos ; e1 F ; e 2 Fn

one can write the following differential equations:

d q
F
dT
( q sin F c T );
( cos n )c T ;
ds
ds T
T

(32)
dx
dz
dS
c T cos ;
c T sin ;
cT
ds
ds
ds

The non-linear differential Eqs. (32) are defined in the global coordinate system xOz (Figure 1), where s is cable
arc (for the inextensible cable) at the current point, as measured from point O; S is cable arc (for the elastic cable)
at the current point, as measured from the same point O; q is the weight (in water) of the unit cable length; x, z are
the current point coordinates;
183

International Journal of Applied Science and Technology

Vol. 4 No. 2; March 2014

T is the cable tension at the current point; is the angle between the x-axis and the tangent unit vector at the
current point; and F
Fp are the tangential and normal components of the hydrodynamic force per unit length of
the cable (Figure 2a). The expressions for both these components are the following:
1
1
w d e c v(z ) cos 2 v(z ) ; Fn w d e c n v(z ) sin v(z ) sin
2
2
where w is the water density; d e is the characteristic diameter of the cable; c
F

(33)

cn are the hydrodynamic

coefficients (c 0 cn = 1); and v( z ) is the flow velocity corresponding to the coordinate z of the cables
current point. Due to the infinitesimal length of the line element, ds, it is assumed that the water flow velocity is
uniform along the element, and it is equal to the flow velocity that corresponds to the z-coordinate of the cables
current point, M (see Figure 2).
The solutions for Eq. (30) and Eq. (33) are obtained by numerical integration after they are transformed in an
algebraic system Eq. (34) and Eq. (35) - by using the finite-difference method with h step size.
2.1- For the Cable Motion in the Plane Xoz:
T j ( t t ) T j1 (t t ) h (a 1, j1 he1, j1 )c T

j (t t ) j1 ( t t )

h
[a 2, j1 e 2, j1 ]c T
T j1

(34)

x j ( t t ) x j1 (t t ) h c T cos j1 (t t )
z j (t t ) z j1 (t t ) h c T sin j1 ( t t )
S j ( t t ) S j1 ( t t ) h c T

Where
a x xi, t z zi, t cos j1 ( t )xi, t sin j1 ( t )zi, t
a n n x x i, t n z zi, t sin j1 ( t )x i, t cos j1 ( t )zi, t
a 1, j1 q cos j1 ( t ) c (x i, t sin j1 (t ) zi, t cos j1 (t ));
a 2, j1 q sin j1 ( t ) c (x i,t cos j1 ( t ) zi, t sin j1 (t ))

e1, j1 F a ; e 2, j1 Fn n a n ; c T 1

T j1 ( t t )
EA

After determining the dynamic configuration, one can find the velocities and the accelerations:
x j ( t t ) x j (t )

x j ( t t ) x j ( t )
t

; z j ( t t ) z j (t )

z j ( t t ) z j ( t )
t

(35)
x ( t t ) x j ( t )
z j ( t t ) z j ( t )
x j ( t t ) x j (t ) j
; z j ( t t ) z j (t )
t
t

j=1, 2, ., n-1
Because the motion of the upper end - point A, (Figure 1) - is imposed, then the last connecting point for
determining the velocity and the acceleration is the connecting point with the index n-1.
2.2- For the Cable Equilibrium Configuration in the Plane Xoz:
184

 Center for Promoting Ideas, USA

www.ijastnet.com

d q
F
dT
( q sin F c T );
( cos n )c T ;
ds
ds T
T

(36)
dx
dz
dS
c T cos ;
c T sin ;
cT
ds
ds
ds

And one can obtain particular solutions for diverse cases, each case being characterized by specific boundary
conditions.
2.1. Moored Systems
2.1.1 Static Analysis
The numerical computation for static analysis is described in the paper [12]. This analysis refers to the both cases,
cable tangent to the bottom and cable detached from the bottom. For all the cases, we have a problem with
boundary conditions at the both ends of cable (bi-local boundary condition).
2.1.2 DynamicAnalysis
The numerical computation begins from the equilibrium cable configuration as the initial mechanical cable state.
a. For the cable of a moored system (as shown in Figure1). The initial mechanical state of cable: initial cable
configuration (the coordinates x, z of cable points in the coordinates system xOz), the tensions in all cable
elements, the angles for all elements; the motion of the upper end (point A): x= x A ( t) , z= z A ( t) ; the velocities
and the accelerations of cable points. The cable is composed by a number of elements. For the elements that are
placed on the sea bottom the tensions are the same and equal to the tension of the first element detached from the
sea bottom.
b. The dynamic cable behavior is determined by solving the differential equations of cable motion (36). For
determining the dynamic response to the excitation induced in the upper end the solution of the differential
equations are determined on the steps of time, t. In the frame of each time step an iterative method is used for
determining the cable dynamic configuration that must satisfy the boundary condition. These conditions can be
attained by fitting the initial value of the cable tension, TA , in the first element that is connected in the upper end
A, and also by fitting the initial value of the angle A of the same element.
The imposed boundary conditions are:
1. In the point a (upper end): x A = x A ( t) , z A = z A ( t) ; the angle A and the tension TA are not known.
2. In the point D-for the cable tangent to the bottom, or the point O for the cable detached from the bottom(lower end): z=0 is known.
3. The cable length is L.
The iterative process begin, first iteration, by fitting the value of the angle A or of the tension TA as follows:
o

A = k1 t , TA = k1 Ta , where 1 0; for the first iteration 1 =1 and t =90 , Ta -the admissible cable
tension; for the second iteration, k=1, 1 =0,5 and k 1 0 , k=1, 1 =0,5 si k 1 0 , k=1,, n max , where n max

is the maximum number of iterations.

The computation stops if the boundary conditions are attained at the lower end, as follows:
At the time step -th:
x A, ( t t ) = x A ( t ) + x A (t t ) = x A ( t ) + a x (sin ( t t ) sin t ) and
z A, ( t t ) = z A ( t ) + z A (t t ) = z A ( t ) + a z (sin ( t t ) sin t );
Where the initial values of the coordinates of upper end A are the values that result for equilibrium configuration,
x A (t 0 ) = x A , z A (t 0 ) = z A
If the point A is connected to a fixed body, for example a quay, and the motion of the moored vessel under the
external excitations is known, x A = x A ( t) , z A = z A ( t) , the computation follows the same steps and the boundary
condition at the lower-point A-are: x n x m ; z n z m where x m and z m are the imposed value of the
coordinates of the point A .
185

International Journal of Applied Science and Technology

Vol. 4 No. 2; March 2014

2.2. Towed Systems

2.2.1Static Analysis
a. The method that was presented for a moored system is also applied for any towed system floating body,
cable or towing vessel. The coordinate system xOz is attached to the towed body in translational motion (Figure
3). In the cables attachment point to the body, O, we have the following conditions: x = z = s = S = 0,
. In
the cables attachment point to the towing ship (point A), we have one only stop condition: either the length of the
cable, L, or another condition that depends on the engineering requirements. For example, should the stop
condition be x = d0, where d0 is an imposed horizontal spacing between the body and the towing ship, or an
imposed value of the depth of the towed body, z= H A . In this case we need to take into account the velocity of
the towing ship in the relationship (7a) that is now:
v r (z i ) [ [v c (z i ) sin ] v x (z i ) v TS ] i [ [ v c (z i ) cos ] v z (z i )]k

where v TS is the velocity of the towing ship as is shown in Fig. 3 for the underwater towed vehicle.
The tension in the attachment point of the cable, To , is understated (this tension is due to the hydrodynamic forces
acting on the body combined with the vehicles weight in water). Also, is known the angle o .
The iterative process begin from the point O and the computation stops if the boundary conditions are attained at
the upper end: if z= H A . The another boundary conditions for stops computation could be: if x= d0 or if s=L
In these cases, we have a problem with boundary conditions at only one of the cable ends.
2.2.2 Dynamic Analysis
A complete dynamic analysis takes into account the motion of all three components of the system: towed vessel,
cable and towing vessel. It is a complex problem that not is our preoccupation in this paper.
However, it must underline that our proposed model for cable dynamics can be used for all cases above presented.

3.The Compound Marine Cables

A typical compound marine cable is made of several components, and one can use the mathematical model that
was established for single cable for any compound marine cables (Figure 4). However, the conditions of
continuity of the compound marine cables must be specified since the boundary conditions are unknown at the
points where the components are located. Moreover, the conditions of continuity in the joints between the
elements must be satisfied. The starting point for the numerical computation can be any one of both compound
marine cables ends and must be re-started at each joint. The static analysis is described in the paper [12].
The dynamic behavior of the cable is determined by solving the differential equations of motion i.e., Eq. (35). For
determining the dynamic response to the excitation induced in the upper end, the solution of the differential
equations is determined along the time, considering the time step amplitude t, and an iterative method is used at
each instant for determining the riser configuration that satisfies the boundary conditions. These conditions can be
attained by fitting the initial values of the cable tension, To , and of the angle o in the element that is located at
the lower end O.
The imposed boundary conditions are:
- At point O (lower end): x=z=0; the angle o and the tension To are unknown.
- At point A (upper end): x= x A ( t) , and z= z A ( t) .
- The buoyancy, B, and length of both parts of the cable, L1, L2 are also imposed.
The iterative process begins by fitting the value of the angle o or of the tension To as follows:
- for the first iteration: o k t , To = k Ta , where 1 0; 1 =1 and t =90o
- for the second iteration: o k t , To = k1 Ta , k=1, 1 =0,5 si k 1 0 , k=1,, n max .
The computation stops if the boundary conditions are attained at the upper end, as follows:
At the -th time step:
For Section 1 (between the point O and point OB):
186

 Center for Promoting Ideas, USA

www.ijastnet.com

- Position of point A at the time step -th:

x A, ( t t ) = x A ( t ) + x A (t t ) = x A ( t ) + a x (sin ( t t ) sin t ) and
z A, ( t t ) = z A ( t ) + z A (t t ) = z A ( t ) + a z (sin ( t t ) sin t );
where the initial values of the coordinates of upper end A are the values that result for equilibrium configuration,
x A (t 0 ) = x w , z A (t 0 ) = H A
- The cable motion is computed by using an iterative method for solving Eq. (35), where the tangential and
normal components of the hydrodynamic force per unit length of the cable are
determined through Eq. (34).
The conditions for re-starting (Figure 5) the computation are: if s= L1 , then

A D h ; FH 1 A x (z ) x (z ) ; 1 c D 2 ; 1 D 2
f f
x
xf

fn
w inf
f
w
f
f
xf
2
2
2

T sin (m f fn )x FH x
1 arctg
B (m f f )z T cos

B (m f f )z T cos T sin (m f fn )x FH x

T 1
cos 1
sin 1

(37)

Where m f is the mass of the buoy, and f and fn are the tangential and normal added mass of the buoy.
For Section 2 (from point OB to point A - upper end):
The numerical computation begins at the re-starting point, OB (see Figure 6).
The following boundary conditions are used
s L1 ; 1 / 2; T T 1

(38)

and the computation stops if s= L1 L 2 and x= x n , (t t ) , z= z n , (t t ) .

If the number of iterations, n, becomes equal to the maximum number nmax and the stop computation condition is
not accomplished, then the solution is not valid. In this case, the boundary conditions must be revised.
For the integration of the Eq. (32) and Eq. (33) we have used the following optimizing procedure:
a) Valid solutions
The solution that corresponds to the pair (T (1), (1)) is valid if it satisfies all the conditions but x (n ) x w and z(n)
= HA. The errors (or conditions) that make a solution not valid are the following:
1. z(j) > HA at Section 1;
2. (j) < 0 at Section 1;
3. (j) > 90 at Section 1;
4. buoy > 0, where buoy is the angle of the cable just after the buoy (at the right side of the buoy);
5. (j) > 0 ;
6. z(j) < 0 at Section 2;
7. (j) > 90 at Section 2;
8. (j) < (j-1) at Section 2.
The variables were scalled as follows: T (1) T.Tad , (1) .90

(39)

Where
T [0,1], [0,1]

(40)

b) The system of equations

187

International Journal of Applied Science and Technology

Vol. 4 No. 2; March 2014

The next step for the admissible solutions (T (1), (1)) or ( T, ) is to solve the system of equations

x (n ) x w ;

z (n ) H A

(41)

that is expressed in the scaled form through

f1 ( T , )

x (n) x w
z (n) H A
0 ; f 2 ( T, )
xw
HA

(42)

and the following four inequations must be added:

0 T 1; 0 1

(43)

It is very difficult to use a standard algorithm to solve these in equations. Hence, the problem is reformulated as
an optimization problem with simple marginal restrictions. This means that the following objective function
should be minimized:
g

1 2
(f1 f 22 )
2

(44)

By taking into consideration the restrictions of simple margin expressed by Eq. (43)
A few methods can be used for solving the optimization problem:
- A version of the quasi-Newton (Broyden-Fletcher-Goldfarb-Shanno, BFGS) algorithm especially adapted to
consider the restrictions. The gradient of the function, g, is computed through finite differences.
- A version of the Newton algorithm where the Hessian matrix is aproximated by finite differences, but modified
in order to consider the restrictions. Additionally, the finite differences approximation of the first order
derivatives is attained with a higher order of precision.
- A Levenberg-Marquardt least squares technique where the Jacobian matrix is approximated through finite
differences.
The aforementioned algorithms were used to solve the non-linear problem represented by Eq. (44) with several
different starting points, always leading to the same solution. This proves that the solution for Eq. (44) is unique.
The method for determining the dynamic behaviour of the cable can be summarized as follows:
i. The initial configuration of the cable - corresponding to the equilibrium configuration is determined according
to the method that is described in reference [12] and detailed above for different cases. Eq. (33) is used for
analyzing the bi-dimensional configurations that are discussed in the present paper.
ii. The cable motion is initiated by applying an external excitation, and the starting point is the static equilibrium
configuration. In most cases the excitation is defined as the displacement, the velocity and the acceleration of the
upper end. The dynamic response of the cable is attained by integration of Eq. (32), and one should notice that
this is done through a computational procedure that is similar to the one that is used for determining the
equilibrium configuration of the cable. This is due to the fact that Eq. (32) is of the same kind of Eq. (33), which
relates to the dynamic equilibrium equation, in which the boundary and the continuity conditions must be satisfied
at each instant of the cable motion.

4.Computed Results
Case Studies
An original computer program (DYNCAB) based on the above-mentioned mathematical model was written to
implement the numerical computation of the complete analysis - static equilibrium configuration and dynamic
behavior of marine cables. An analysis of the static equilibrium configuration by using the abovementioned
program, as well as the validity of the approach, was made by Matulea et al. [12]. This was done by comparing
the DYNCAB results with the results obtained through a lamped mass method [13]. Additionally, Matulea et al.
[12] have shown that there is a good correlation between the result of numerical simulations and the results
attained with the tank tests. The present paper shows that the approach is also suitable for a comprehensive
dynamic analysis in order to find the best solution for different operating conditions.
Case study: Model of a compound cable with a float
Now, our goal is to validate the proposed approach for the dynamic behaviour of a marine cable.
188

 Center for Promoting Ideas, USA

www.ijastnet.com

Figure 8a and 8b show the results obtained by using the proposed method. For a matter of comparison, the results
obtained through the lamped mass method of Nakajima et.al [13] are also shown. As one can see, there is a
good correlation between the proposed method and the lamped-mass method.
The values of the vertical component of the cable tension in the upper end A, TA,z , are the same for both the
proposed method and lumped-mass method. The same apply to the experimental tests as well (see Figure 8a
and 8b). In the Figure 8a there are noted by TA0 , x and TA0 ,z the initial values of the tension components on
coordinate axes in the upper end A. These values result from static analysis.

5. Concluding Remarks
A novel method based on a very flexible and efficient mathematical model for the complete static and dynamic
analysis of marine cable dynamics was presented in this paper. The model takes into account most of the nonlinearities of marine cable, such as nonlinear water velocity profiles, large marine cable deflections and nonlinear
constraints.
The computer program that implements the method is a general purpose one. It can handle the instability that is
due to bifurcation, and makes it possible to study a large number of operating conditions and system
configurations in a cost-effective way.
Using the program for analyzing some cases of marine cable tested the soundness of the model. The numerical
experiments that are presented in the paper show a very good correlation with the results obtained by Nakajima et
al. [13] as shown in Figures 7a and 7b.
The use of lumped mass models imparts some practical difficulties: first, the marine cable length must be
provided as input data; second, the data preparation is a laborious task; third, it is hard to find out the location of
the point where the marine cable contacts with the seabed.
On the contrary, the mathematical model hereby presented proved to be more flexible and efficient because it is
based in non-restrictive assumptions that express very well the most typical engineering design requirements. In
fact, the new model allows for the use of the cable arch as an integration variable, as well as the marine cable
coordinates. In addition, the proposed model makes very easy to consider different boundary conditions as well as
distinct continuity conditions.
As a general conclusion arising both from the theoretical standpoint and the results of the performed
calculations - one can point out:
- static analysis is particularly useful in defining the design parameters and the initial configuration that is
required to perform the dynamic analysis;
- dynamic analysis is required to evaluate the occurrence of large stresses due to dynamic loading, and the
analysis in the time domain allows a very complete and accurate prediction of all the most important aspects of
the marine cables behavior.
In order to improve the computer program, it is still necessary to add the possibility of taking into account the
variable marine cable bending stiffness and the variable buoyancy close to the sea surface, as a means to allow the
analysis of some more involved marine cables systems that are being studied by some of the program users.

Acknowledgements
The present work was partially supported by S.N. PETROM S.A., branch office PETROMAR, Constana,
PETROSTAR S.A., Ploieti, Naval Research Center, Constana, National Ministry of Education and Research of
Romania (through CNCSIS - National University Research Council) and Ministry of National Defense, under
several R&D contracts. This support is gratefully acknowledged.

189

International Journal of Applied Science and Technology

Vol. 4 No. 2; March 2014

References
Abkowitz, A.M., (1969), Stability and Motion Control of Ocean Vehicles. The MIT Press.
Ablow C.M., Schetcher, S., (1983), Numerical simulations of undersea cables dynamics. Ocean Engineering,
10(6): 443-457.
Bernitsas, M. M., (1982), A three dimensional nonlinear large deflection model for dynamic behaviour of risers,
pipelines and cables. Journal of Ship Research, 26(1): 59-64.
Berteaux, H.O., (1976), Buoy engineering. John Wiley and Sons, N.Y.
van den Boom, H.J.J., Dekker, J.N., van Elsaker, A.W., (1987), Dynamic aspects of offshore riser and mooring
concepts. In: Proc. 19th Annual Offshore Technology Conference, paper OTC 5531, Houston.
Chapman, D.A. , (1984), Towed cable behaviour during ship turning manoeuvres. Ocean Engineering, 11(4): 327361.
Chatjigeorgiou, I.K., (2008), A finite difference formulation for the linear and nonlinear dynamics of 2D catenary
risers. Ocean Engineering, 35: 616-636
Hosseini Kordkheili, S.A.,Bahai, H., (2008), Non-linear finite element analysis of flexible risers in presence of
buoyancy force and seabed interaction boundary condition. Arch. Appl. Mech., 78:765774.
Larsen, C.M., (1992), Flexible riser analysis-comparison of results from computer programs. Marine Structure, 5,
107119.
Low, Y.M. and Langley, R.S. (2006b). Dynamic analysis of a flexible hanging riser in
the time and frequency domain, OMAE 2006-92171, Hamburg, Germany.
Matulea, I., Talmaciu, N., (2000), Static and dynamic analysis of flexible system for hydrocarbon transfer. In: Proc.
5th International Conference on Marine Science and Technology, pp. 43-47, Varna, 2000.
I. Matulea, A. Nstase, N. Tlmaciu, G. Slmnoiu, A. M. Gonalves-Coelho, (2008). On the equilibrium
configuration of mooring/towing lines. Applied Ocean Research, 2008; 39(2): 81-91.
Nakajima, T., Motora, S., Fujima, N. , (1982). On the dynamic analysis of multicomponent mooring lines. In:
Proc. 14th Annual Offshore Technology Conference, paper OTC 4309, Houston.
Rodrigues, G.J.O., Cardoso, D.C.T.,deLima,B.S.L.P., Jacob,B.P., Fernandes,A.C., (2005).
Ananalytical/numerical procedure for structural analysis of hybrid riser systems. Proceedings of the
24thInternational Conference on Offshore Mechanics andA rctic Engineering,vol.1;2005,pp.549555.
Santillan, S.T.,Virgin, L.N.,Plaut,R.H., 2008. Nonlinear two-dimensional elastic analysis of subsea
risers. In:Proceedings of EURODYN2008, 7th European Conference on Structural Dynamics Paper E122.
Santillan, S.T.,Virgin, L.N., Plaut,R.H., 2010. Static and dynamic behavior of highly deformed
risers and pipelines. J.Offshore Mech. Arctic Eng.132.
Triantafyllou, M.S. , (1982). Preliminary design of mooring systems. Journal of Ship Research, 1982; 26(1): 2535.

Nomenclature
- t - virtual angle used in computation process that must be found by trial and error.
- - angle between the tangent to the riser and the z-axis at the current point
- - angle between the unit vector and the vector of the current water velocity at coordinate z
- v(z ) - relative velocity to the water
- vc(z) - water current velocity
- w - water density
- (rv) - vector of the current position, at the time t, of the point M
- (r ,t ) - velocity vector of the point M at the time t
- (r ,t ) - acceleration vector of the point M at the time t
- Fn, Ft - normal and tangential components of hydrodynamic forces acting on the unit cable length
- (F ) - added mass vector
- , n - added mass for a unit length of cable
- c in , c i - inertial coefficients
- A - effective area of the cable section
- cn, c - coefficients of hydrodynamic forces
190

 Center for Promoting Ideas, USA

www.ijastnet.com

- dc - characteristic diameter of the riser

- E - longitudinal modulus of elasticity
- [R] - transform matrix between the system of vectors , n and the coordinate system xOz
- h - computational step in the finite difference method
- H - water depth
- q - line weight in water per unit length
- s - arc of non-stretched line at current point
- S - arc of the stretched line at current point
- T - cable tension in the current point
- x, z - coordinates of the current point
- B - buoyancy
- L1, L2 - lengths of the components of a multicomponent line
- xA, zA - coordinates of the upper end

Figure 1 Mooring Cable

Figure 2. The Forces Acting on an Element of the Cable

191

International Journal of Applied Science and Technology

Figure 3 Towing Cable

Figure 4 Mooring Cable with a Float

Figure 5. The dynamic Loading of the Buoy

Figure 6 Case study: Cable with a Buoy

192

Vol. 4 No. 2; March 2014

 Center for Promoting Ideas, USA

www.ijastnet.com
Table 1-Cable Characteristic

Material
Weight in air
Weight in water
Volume/unit length
Elasticity modulus (E)
Diameter d

N/m
N/m
cm3/m
N/m2
cm

Steel
2,22
1,938
28,2
21,5x1010
0,599

1
5
10
14
18
19

0,00
4,53
9,00
13,10
16,10
18,00

46,2
49,5
47,4
47,5
48,2
50,35

0,00
0,00
15,30
20,25
22,34
24,00

0,00
4,11
9,04
13,02
16,52
17,60

(
m
)

(
d
e
g
x

T
e

(
s
i
o
n

"
s
"

n
oc

o
i
n
t

Table 3-Characteristic of Equilibrium Configuration

(
m
)

Table 2-Buoy
0,7 N
-30 N

Weight in air
Weight in water

0,00
0,9
2,35
2,39
2,73
3,00

Figure 7a: The Model Behavior under Dynamic Loading

193

International Journal of Applied Science and Technology

Vol. 4 No. 2; March 2014

Figure 7b: The Model Behavior under Dynamic Loading for the Time Period T=1.2 S

194